Predicted flight trajectories differ depending on which orientation cues are used by migrating birds. Results from radar and satellite tracking of migrating birds can be used to test which of the predicted trajectories shows the best fit with observed flight routes, supporting the use of the associated orientation mechanism. Radar studies of bird migration at the Northeast Passage and the Northwest Passage support the occurrence of migration along sun compass routes in these polar regions. In contrast, satellite tracking of Brent geese (Branta bernicla) migrating from Iceland across Greenland and from Northwest Europe to Siberia show routes that conform most closely with geographic loxodromes, but which are also profoundly influenced by large-scale topography. These evaluations are discussed in relation to the adaptive values of different routes in different parts of the world. Sun compass routes are favourable mainly for east-west migration at high latitudes. For east-west migration at mid and high latitudes magnetic loxodromes are more favourable than geographic loxodromes in certain regions while the reverse holds in other regions. The geometry of migration routes, as recorded by radar and satellite tracking, may be important for understanding the evolution of the complexity of birds' orientation systems, and for providing clues about the orientation mechanisms guiding the birds on their global journeys.